Vertical Suspended Sediment Concentration Due To Waves And Wave-current Flows

According to the mechanism of sediment suspention, the suspended sedimentconcentration profiles due to waves and wave-current flows are analyzed. Themathematical models are set up respectively. Under waves the motion of sediment isnot so complicated relatively that the explicit formulae can be derived. However inwave-current flows the motion is so complicated relatively that implicit solutions isneeded because that the explicit solutions are not precise enough. The implicitsolutions can be gained through an iterative program. The sea bed is very rarely flat inreal sea. It tends to be covered by sediment structures such as regular and irregularripples, which will influence the suspended sediment concentration significantly. So itis important to predict the shape of the ripple precisely. Here a new method, ArtificialNeural Network, is used to predict the length and height of ripples. The details of thestudy and conclusions are as follows:1. A feed forward network with back propagation of errors is used to predict thelength and height of a ripple. It is trained based on a large number of data atlaboratory and in field. The trained network is proved to be valid by comparisonbetween the prediction and the measurement.The network is trained only based on thedata of ripples under waves. However if the current is weak, the current will notinfluence the shape of ripples significantly, for which the network can still be valid insuch a condition.2. It is appropriate to use a layered model to describe the distribution ofsuspended sediment concentration due to waves. In the boundary layer the diffusioncoefficient is gained based on the turbulent diffusion;in the heigher flow layer thediffusion coefficient is obtained based on a mixing length theory;between the abovetwo layers there is a linear transition of diffusion. The concentratin is gained bysolving the diffusion eqution. The values of the suspended sediment concentrationcomputed are in satisfactory agreement with data from laboratory and field.3. A continuous diffusion model is derived from a continuous eddy viscositymodel with the help of the steady parameter. The concentration and the steadyparameter are calculated in an iterative program until both of them converge. Thevalues of the suspended sediment concentration computed agree satisfactorily with thefield's measurements.